The impact of various medium conditions on cellular proliferation/viability was analyzed using SRB assay as described in great detail by (Vichai and Kirtikara 2006). To
determine changes of cellular proliferation upon incubation with various medium conditions, approximately 2 × 104 cells were seeded in full medium in 48-plate format, 24 h prior to the treatment. Cells were then washed once with PBS and fed with the chosen medium. Cell proliferation/viability was analyzed at different incubation time points using SRB assay as described in great detail by Vichai and Kirtikara (Vanicha Vichai 2006, Nature Protocol). In brief, cells were fixed with ice-cold 10% (trichloroacetic acid) TCA for 1 h; fixed cells were then stained with 0.054% w/v SRB for 30 min and
absorbance was measured at 535 nm using a Tecan Ultra plate reader (Tecan, Maennedorf, Switzerland).
10. Results
10.1. MYCN promotes cell cycle entry
10.1.1. Reduction of MYCN leads to prolonged doubling time
Previous studies have found that MYCN boost proliferation(Meyer and Penn 2008). Here we used a tet inducible shMYCN cell line, IMR5/75, that has 75 copies of the MYCN oncogene. The reduction of MYCN is not complete (Figure 11A) but significant. Through the text, MYCN-low cells will be denoted as the cells were MYCN is reduced by the shRNA. There were not reports of doubling times with respect to this particular cell line. To get accurate estimates of doubling times, IMR5/75 cells were counted by FACS over a period of a week to calculate the growth curve (Figure 9B). This result is necessary for further calculations of cell cycle phase lengths that were performed in the further sections. To confirm this result, cell growth was additionally measured by impedance. IMR5/75 cells either with or without the E2F-d2GFP construct were seeded in the RTCA plates. These device measures electron flow on the surface of the plate. Attached cells will interfere with the electron flow generating impedance The higher the cell density the higher the impedance, therefore serving as a readout of cell proliferation. Impedance in some situations could also indicate changes in cell attachment. Overall the results are shown in Figure 9A. The data shows strong accordance between both types of measurements as shown in Figure 9C. MYCN-high cells exhibit a doubling time of around 16-21 hrs and MYCN-low cells between 35-50 hrs. This clearly indicates that reduction of MYCN almost increase by half the length of the cell cycle duration.
Figure 9 Growth kinetics of MYCN amplified IMR5/75 cells. A) Impedance measurement of cell growth.
2 biological duplicates are shown for MYCN-high/low condition. B) Growth curve calculated by counting viable cells. SEM is calculated from 3 biological replicates. C) Comparison of doubling times calculated
Also interesting to note, MYCN-high cells do not remain viable for longer than 6 days as shown in the impedance data (Figure 9A). Perhaps in confluent conditions ROS levels are overly toxic for active proliferating cells as shown later in the metabolic section. Calculation of the doubling time is mathematical measure that generalizes a value based on a population measure. To test if this correlates with the actual growth of the cells, MYCN-high cells as an example were sorted in 6-well plates based on G1 DNA content and low E2F-d2GFP reporter (Figure 14A). This reporter cell line previously generated in the Master thesis will be used here in combination with Cdt1-degron marker for characterizing the restriction point in these cells. The results show that after
50 100 150 200 250 -1 0 1 2 3 4 5 Time (hr) Ce ll in d e x
Cell index = A* 2^(t/CI doubling-time)
MYCN-high (1st duplicate) MYCN-high (2nd duplicate) MYCN-low (1st duplicate) MYCN-low (2nd duplicate) 0 24 48 72 96 120 0 2×105 4×105 6×105 8×105 1×106 Time (hr) (c el ls /m l) MYCN-high MYCN-low A B C
Impedance Growth curve 0 20 40 60 Do ub lin g tim e (h r) MYCN-high MYCN-low
15hr cells populate all the phases, seemingly correlated with the doubling time calculated based on growth curves and impedance (Figure 10). Interestingly, cells from low E2F-d2GFP levels are still able to grow in the MYCN-high condition. This is confirmed with further experiments later in this thesis, showing that MYCN-high cells do not seem to have a permanent quiescence phenotype.
Figure 10 Cell sorting outgrowth. Left panel: MYCN-high cells sorted by low E2F1 promoter activity in
G1. Right panel: Measurement of sorted population after 15 hr incubation.
In this particular context, it is important that MYCN is the main driver of the observed phenotype. However the effects can be masked by its homologous c-MYC. However FACS protein staining of c-MYC do not show significant differences upon MYCN reduction (Figure 11A: right panel). Therefore it is safe to say that MYCN is the main driver of the cell cycle effects observed in this particular cell line. The growth measurements showed that MYCN reduction lengthens cell cycle duration by almost a factor of 2. To explore if this is can be explained by the proportion of G0 cells, EdU incorporation was measured after a 30 min pulse. The results indicate that reduction of MYCN increase the proportion of dormant or non-proliferating cells by 10% (Figure 11C). This is still a modest increase but shows that MYCN might be promoting proliferation by inhibiting G1 arrest.
15 hr post-sort 0 200 400 600 800 1.0K DNA content 100 101 102 103 104 E2F-d2GFP (log 10 a.u) 0 200 400 600 800 1.0K DNA content 100 101 102 103 104
10.1.2. MYCN favors crossing of the restriction point
The Rb-E2F network mostly regulates G1/S transition. The way the mechanism works is by deactivating Rb via phosphorylation leading to E2F1 release from the repressing complexes and increase of E2F1 expression via auto-activation. To explore if the effects observed in the proliferative and dormant fractions are due to changes in the most important G1 regulators (Figure 11B), protein staining of pRb807/811 phosphorylation site (a marker for cell proliferation) E2F1 total protein and d2GFP we co-stained with DNA content staining using FxCycle violet dye. Gates for the cell cycle phases were applied using the Dean Jet Fox model in FlowJo. For each of the cell cycle phases, the distributions were plotted with either the autofluorescence control or the isotype staining. As previously found with EdU incorporation assay, MYCN-low cells accumulate in G1 with almost 10% more cells. The distributions show some interesting behavior. pRb exhibits a bimodal behavior in G1, with almost unchanged levels in the other cell cycle phases. MYCN-low cells exhibit a greater accumulation in the negative pRb fraction as compared to MYCN-high. This shows that MYCN-high cells do maintain a greater proportion of cells in the phosphorylated state, probably the proliferating cells. On the other hand, E2F promoter activity (E2F-d2GFP) exhibits a strong shift in G1 with moderate shifts in the other phases, favoring usually higher promoter activity levels in MYCN-high cells. E2F1 protein levels remain slightly changed in G1 mostly.
Figure 11 Characterization of the cell system. A) Antibody staining shows MYCN reduction upon
shRNA induction. c-MYC levels remain unchanged. B) pRb807/811 and E2F1 staining together with E2F-
A
100 102 104
EdU (log10 a.u) 0 20 40 60 80 100 100 102 104
EdU (log10 a.u)
% Cells MYCN-high MYCN-low 60.3% 39.7% 72.7% 27.3% 0 20 40 60 80 100 0 20 40 60 80 100 102 103 104 0 20 40 60 80 100 102 103 104 102 103 104 E2F1 (log10 a.u)
E2F-d2GFP (log10 a.u)
pRb807/81 1 (log10 a.u) G1/0 S G2 MYCN-high: 56.5 % MYCN-low: 67.4 % 29.0 %20.1 % 14.2 % 11.9 % MYCN-high MYCN-low
B
C
102 103 104 0 20 40 60 80 100 % Cells 102 103 104
d2GFP activity for each cell cycle phase. Faded gray distributions indicate either auto-fluorescence control or isotype. C) EdU incorporation assay. The gate was calculated with the bifurcate option in FlowJo.
Together this indicates that G1 key regulators are strongly affected by MYCN. An interesting point is that E2F1 promoter activity did not show a bimodal behavior as previously reported by (Yao, Lee et al. 2008). However pRb shows a clear bimodality. The relevance of this finding is that its commonly assumed that cell cycle checkpoints work as bistable switches exhibiting ultrasensitive behavior, that means if titration of a signal is performed, the response curve will look sigmoidal, since low amount of signals do not trigger response only until certain threshold is crossed. In single cells measures this ultrasensitivity is observed by bimodal distributions that switch from negative to positive stats without intermediates. This suggest that the bistable switch might be still intact in MYCN-high cells but MYCN might increase the probability of crossing the switch evidenced by a higher proportion of pRb positive cells. As for the case of the the E2F promoter, perhaps external sources of noise could mask the bimodal behavior, but further experiments are required to clarify the noise role in MYC-driven tumors. To investigate further the integrity of the bistable switch, we took advantage of normal heterogeneity of protein levels in cells. We performed a dose response curve of IMR5/75 cells stained with MYCN antibody as the signal and then we looked at the E2F-d2GFP levels of those cells. The MYCN protein distribution was sliced by intervals of 5%, covering all possible MYCN levels using a custom in-house script in FlowJo (Figure 12 left panel). The median values from the obtained distributions after slicing were computed for from the E2F promoter construct and plotted against the median MYCN levels as shown in Figure 12 (right panel). The E2F promoter construct values were normalized to 1 to allow the behavior to be comparable between MYCN-high and low cells. Each point in the dose response curve was mapped to the cell cycle phase using the median values for DNA content. Hill equation was fitted to the data using the least squares method. Interestingly, MYCN-high cells exhibited hyperbolic behavior, typically found in Michaelis-Menten type systems and denoted here by hill coefficient of 1. In contrast, MYCN-low cells showed ultransensitive behavior with hill coefficient of 4, strongly suggesting the existence of a bistable switch as also evindeced by the bimodal
behavior of pRb. Our results suggest that oncogenes might favor crossing of the bistable switches by allowing low input signals to engage directly into the high activation state, in this case the transition from G1 to S phase.
Figure 12 MYCN dose response curve. Left panel: MYCN distributions were sliced by 5% percentiles
and the intnsity values were plot against E2F-d2GFP values. Middle panel: MYCN-high dose response exhibits hyperbolic behavior. Right panel: MYCN-low dose response exhibiting sigmodail behavior.
10.2. MYCN drives scape from G0-like state
10.2.1. Cdt1-degron and E2F1-d2GFP reporters define G1 sub-phases
This indicates that MYCN-high cells are perhaps more sensitive to growth signals that MYCN-low cells. A higher probability to cross the restriction point might be due to relatively faster transition or skipping transiently arrested states such as G0. Up to now, only few markers are available that could describe the arrested states in a population of single cells. For that purpose, the previously generated cell line with the E2F-d2GFP promoter construct was transfected with the Cdt1-degron construct which is a marker for G1 length. The schematic of the constructs is depicted in Figure 14A. The Cdt1-degron construct has a Cdt1 fragment that contains a SKP2 degradation site. An example of a microscopic trajectory for Cdt1-degron and E2F-d2GFP dynamics is show in Figure 14B. At the beginning of the cell cycle or birth, the cells are colorless but once cells enter in G1, Cdt1-degron begins to accumulate. Interestingly, this particular MYCN-low cell exhibit higher Cdt1-degron intensity. Previous reports have shown that this might be a proxy for G0 arrest. As for the case of E2F-d2GFP, the patterns look similar between
0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 nh = 1.0 MYCN-high MYCN-low 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 nh 4 Relative E2F-d2GFP In te ns ity ( a. u)
Relative MYCN Intensity (a.u)
G1 S/G2 G1 S/G2
% Cells
MYCN (log10 a.u)
MYCN-high
MYCN-low
MYCN-high and low cells with the difference that MYCN-high cells exhibit a higher promoter activity as previously found with the FACS data.
Figure 13 Reporters Cdt1-degron and E2F-d2GFP used to characterized G1 in neuroblastoma cells.
A). Schematic representation of the plasmids used. In blue it is indicated the SKP2 recognition site for the Cdt1-degron reporter. E2F-d2GFP reporter scheme with c-MYC and E2F1 biding sites. B) Example
trajectories which each of the reporters.